Bent plate after welding. How to prevent? 1

[cooperjer]

I am faced with a design problem that involves welding bending a flat plate. The flat plate is 23.5x23.5x.38 –inches. I’m welding a two 3.00x3.00-inch tubes about 9 inches apart along the center line of the plate. In addition I’m welding 2 L brackets near the center of the plate about 3 inches apart. All this welding on one side of the plate is causing the plate to bend, as would be expected. I’m looking for design ideas that will result in a flat plate after welding.

One common solution is to pre-bend the plate in the opposite direction that the weld bends it, however this plate is being used on two sides of a box and is not symmetrical.

Thank you for your replies.


Cooperjer
Mechanical Engineer

 

[BillPSU]

Put your parts through the plate and weld on both sides would be one solution.

Give your weld fixture the ability to clamp your plate into a pre-stressed postion. Trial and error will need to be done to determine the correct pre-stress.

Don't weld, bolt.

Straighten plate after welding.

Machine plate after welding.

Stiffener on plate to resist distortion.

If you closely examine the plate around the 3 x 3 tube you will probably find a localized out of flat condition also. How flat do you want the plate after welding?

 

[desertfox]

Hi cooperjer

Have you considered welding to a thicker plate and then doing a post weld stress relieve followed by machining the plate down to the right thickness.


regards desertfox

 

[MikeHalloran]

Speaking of machining, in some applications it can make sense to just take a real thick plate and mill off what you don't need, even if, maybe especially if, the desired end result is a thin plate with a few features on it.

The only continuous labor cost is for shoveling the chips out of the CNC mill, and some of them do that for you, too.



Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[cooperjer]

"How flat do you want the plate after welding?"
This is a good question. On two corners of the plate a shaft protrudes out opposite of the 3x3 tubes about 6-inches. This shaft has a plastic roller on it. Any deflection in the plate due to welding or pre-welding manufacturing, as the plate has holes in it, causes the corners to bend toward the 3x3 tube side. This bending prevents the roller from riding flat on the rails. The result of course is increased wear on the rollers and rail. Thus far testing has shown a deflection of about in the plate of .06 to .09 inches due to welding. The addition of the weight of the machine on the rollers causes an additional .06 to .09 inches of deflection in the plate. There is no black and white answer to how flat I want the plate.

A post welding process may be a possibility. I’ll check with fabrication and get their opinion.

Thank you for the responses.


Cooperjer
Mechanical Engineer

 

[diamondjim]

I would think you could apply a torch
to the other side to help bring it
back to a nearer flat condition.

 

[metengr]

Better be careful with heat straightening. Nobody has mentioned what material is being used for assembly. If it is low to medium carbon steel plate, I would use a Temp Stick to monitor temperature and stay below 1200 deg F metal temperature during the application of heat. Spot heating may be required to achieve the desired flatness.

I would practice heating on a piece of scrap plate before trying this on your welded assembly.

 

[cooperjer]

Forgot the steel type. It's 304 Stainless.

Cooperjer
Mechanical Engineer

 

[metengr]

ok. This changes the dynamics of the situation. Your best option is to carefully use spot heating to attempt to straighten the plate, I would not heat the entire assembly. Spot heating will require some trial and error to perfect, but it could be done.

 

[unclesyd]

I once asked a guru of straightening steel about straightening SS.
His answer was, has it been welded on?
I answered yes.
His answer was that you had better say a prayer.

You will be able to straighten some SS parts but the very next one will get catty wampus, not telling where it will go.

 

[BillPSU]

Is it possible to use radiused roller to accept the out of parallel condition? Localized roller deformation may give you enough surface contact to support the part.

 

[MikeHalloran]

If normal deflections cause the assembly to malfunction, it's time to re-evaluate the design of the assembly, not find ways to make the components less imperfect.



Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[cooperjer]

Thank you all for the responses. I’ve presented by manager with 5 possible design solutions: 1) pre-weld bending, 2) post weld bending, 3) pre-weld heating, 4) post weld heating / straightening, and 5) added gussets to prevent deflection. Now it’s time to weigh the merits and flaws of each option.

Thank you again.


Cooperjer
Mechanical Engineer

 

[weldtek]

One other suggestion is to use techniques, joint designs and select processes resulting in the lowest possible heat input. This covers a lot of ground and requires some careful consideration.

 

[griffengm]

Cooperjer,
Are the tubes welded continuously along the length of the sheet or are they stitch welded? Stitching can reduce the amount of distortion. Another variation of this is to stitch while alternating sides of the tube.
I have had some experience with heat straightening on mild steel and it is easy to turn "some" distortion into a "lot" so I would be very cautious on a stainless assembly. If this is your first go at this, plan on sacrificing at least a couple before you have consistency. How much are you willing to risk?

Griffy

 

[unclesyd]

Did you mention holes in the bent part of the plate?
If so, how many of what size?
Is the total deflection from welding and operating load .120" and .180"?
How are the L brackets welded. flat side or on edge with flat side up?

One possible way out if your design will allow it cap/plug the end of the tubes and plug weld from the opposite side with whatever stitch welding is needed on the tube side.


A secret trick that may work on the plate if there aren't too many holes.

Assuming this is a mechanical contrivance.
If I understand correctly the plate has bent up leaving the bottom convex. Turn the assembly upside down lay a straight edge across the plate and find just where the plate starts to break away from the straight edge. Get your best light handed tig welder and let him do a very light fusion pass along a line at the point where plate starts to breakaway from the straight edge. The line will probably not be symmetrical. Only weld where there is a break. You might have to do slightly over lapping parallel passes in the worst area.

 

[Hydroformer]

Unclesyd:

I thought about mentioning redundant 'compensating' welds on the opposite side. We try to keep this type of correction as a 'last resort' in production (cycle time, mat'l, added heat / cooling, etc.) But, it does work.

I've even seen on a high volume, large tubular mig-welded assembly, (where multiple fixtures are in use) an inline probe check that provides feedback to a PLC which in turn calls up a "corrective action" and directs a robot to provide a suitable compensatory weld(s) to pull the assembly back towards nominal.

The "corrective action" table was compiled through empirical data collection and a loose DOE. This effort was well rewarded with a dramatic increase in overall CpK ratings.

regards,

Hydroformer

 

[MURZ]

You can also look at your welding procedure. Can you decrease the weld size any? This will help with the shrinkage. (e.g., smaller weld smaller shrinkage, smaller distortion). Angular distortion in fillet welds is affected by over-welding. Look at reducing the total heat input into the part (amps x volts x 60 / travel speed) without sacrificing weld integrity. Less heat input less distortion. Can you use skip welds in lieu of full-length fillet welds (e.g., 2” weld on 4” centers). Are you making more that one of these panels? If so can you fabricate them using a back-to-back assembly, his will give you additional stiffness. You can clamp or tack the assembly together and weld your L and tube to the face of each. If you can tack-weld the plates together you will need to stress relive the components before separating. If you do not want to perform the stress relief, then use clamps and insert wedges between the panels so when the wedges are removed the parts will move back to the correct shape. Also, look at weld sequence. Typically, the direction of should be towards the free end of the joint/weld. Also, use short welds or runs by using the back-step or skip welding technique. These are very effective at controlling distortion.

Hope some of these help.

Mark

 

[MetalsInc]

cooperjer: Any decisions or results?

 

[cooperjer]

We have decided to change the design to bolt the L-bracket to the flat plate. Thank you all for the suggestions. I have considered all of them.

Cooperjer
Mechanical Engineer